ESOM Visualizations for Quality Assessment in Clustering

Ultsch, Alfred; Behnisch, Martin; Lötsch, Jörn

doi:10.1007/978-3-319-28518-4_3

Cited by 9 publications

(19 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Present results reproduce earlier demonstrations that ESOM/U-matrix correctly detects clusters in artificial [14] and biomedical data sets [39], while overcoming the imposing of spurious clusters [12,14]. When the structure in biomedical data is destroyed by permutation, classical clustering algorithms may still suggest a structure.…”

Section: Discussionsupporting

confidence: 87%

Identification of disease-distinct complex biomarker patterns by means of unsupervised machine-learning using an interactive R toolbox (Umatrix)

et al. 2018

Self Cite

View full text Add to dashboard Cite

Background: Unsupervised machine-learned analysis of cluster structures, applied using the emergent self-organizing feature maps (ESOM) combined with the unified distance matrix (U-matrix) has been shown to provide an unbiased method to identify true clusters. It outperforms classical hierarchical clustering algorithms that carry a considerable tendency to produce erroneous results. To facilitate the application of the ESOM/U-matrix method in biomedical research, we introduce the interactive R-based bioinformatics tool "Umatrix", which enables valid identification of a biologically meaningful cluster structure in the data by training a Kohonen-type self-organizing map followed by interface-guided interactive clustering on the emergent U-matrix map. Results: The ability to detect clinical relevant subgroups was applied to a data set comprising plasma concentrations of d = 25 lipid markers including endocannabinoids, lysophosphatidic acids, ceramides and sphingolipids acquired from n = 100 patients with Parkinson's disease and n = 100 controls. Following ESOM training, clear data structures in the high-dimensional data space were observed on the U-matrix, allowing separation of patients from controls almost perfectly. When the data structure was destroyed by Monte-Carlo random resampling, the U-matrix became unstructured and patients and controls were mixed. Obtained results are biologically plausible and supported by empirical evidence of a regulation of several classes of lipids in Parkinson's disease. Conclusions: Sophisticated analysis of structures in biomedical data provides a basis for the mechanistic interpretation of the observations and facilitates subsequent analyses focusing on hypothesis testing. The freely available R library "Umatrix" provides an interactive tool for broader application of unsupervised machine learning on complex biomedical data.

show abstract

Section: Discussionsupporting

confidence: 87%

Identification of disease-distinct complex biomarker patterns by means of unsupervised machine-learning using an interactive R toolbox (Umatrix)

et al. 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the case of self-organizing mapping (SOM) 15 , the structures have been reported to be of "very general shapes" [Duda et al, 2001, p. 582;Ultsch/Lötsch, 2016]. Similarly to the emergent SOM (ESOM)/U-matrix clustering method [Ultsch et al, 2016a], the Databionic swarm (DBS) method that is discussed later in this work also uses the concept of emergence 16 , through which novel properties can arise in a system. Emergence leads to clusters whose structures are not predefined.…”

Section: Types Of Structures Sought By Clustering Algorithmsmentioning

confidence: 99%

“…For example, Cottrell and de Bodt 20 For an overview, see [H. Ritter et al, 1992], for deep learning see [Goodfellow et al, 2016]. used 4x4 units to represent the 150 data points in the Iris data set ( [Ultsch et al, 2016a] cites [Cottrell, 1996]). Therefore, the conventional SOM algorithm is called k-means-SOM here.…”

Section: Emergent Self-organizing Map (Esom)mentioning

confidence: 99%

“…Combining ESOM with the U*-matrix approach enables an application of [Ultsch et al, 2016a]: In contrast to the Databionic swarm approach, in which the shortest paths between AU-distances are calculated 23 , this clustering approach uses only the direct neighborhood of the projected points. A computation of the abstract P-matrix is necessary because ESOM itself does not consider density.…”

Section: Clustering With Esommentioning

confidence: 99%

“…From this starting point, the user may search interactively for the empirical Pareto percentile [that] [Ultsch et al, 2016a]. It can be formalized as [a] pointwise matrix [product] [Thrun et al, 2016a] …”

Section: 11mentioning

confidence: 99%

See 2 more Smart Citations

Projection-Based Clustering through Self-Organization and Swarm Intelligence

Thrun¹

2018

View full text Add to dashboard Cite

Using Projection-Based Clustering to Find Distance- and Density-Based Clusters in High-Dimensional Data

Thrun

Ultsch

2020

J Classif

Self Cite

View full text Add to dashboard Cite

For high-dimensional datasets in which clusters are formed by both distance and density structures (DDS), many clustering algorithms fail to identify these clusters correctly. This is demonstrated for 32 clustering algorithms using a suite of datasets which deliberately pose complex DDS challenges for clustering. In order to improve the structure finding and clustering in high-dimensional DDS datasets, projection-based clustering (PBC) is introduced. The coexistence of projection and clustering allows to explore DDS through a topographic map. This enables to estimate, first, if any cluster tendency exists and, second, the estimation of the number of clusters. A comparison showed that PBC is always able to find the correct cluster structure, while the performance of the best of the 32 clustering algorithms varies depending on the dataset.

show abstract

ESOM Visualizations for Quality Assessment in Clustering

Cited by 9 publications

References 9 publications

Identification of disease-distinct complex biomarker patterns by means of unsupervised machine-learning using an interactive R toolbox (Umatrix)

Identification of disease-distinct complex biomarker patterns by means of unsupervised machine-learning using an interactive R toolbox (Umatrix)

Projection-Based Clustering through Self-Organization and Swarm Intelligence

Using Projection-Based Clustering to Find Distance- and Density-Based Clusters in High-Dimensional Data

Contact Info

Product

Resources

About